Fan Flow Deflection for Supersonic Turbofan Engines
Dimitri Papamoschou∗ and Preben E. Nielsen†
University of California, Irvine, CA, 92697-3975
We present an initial parametric investigation of fan flow deflectors for suppressingnoise from supersonic turbofan engines. Realistic exhaust geometry and flow conditionsfor bypass ratio 2.7 were simulated in a subscale experiment. The study encompassedacoustic measurement and mean velocity surveys. The deflectors comprised internal vaneswith both symmetric and cambered airfoil sections and deployable external flaps. Superioracoustic results were achieved using a combination of cambered vanes and perforated flaps,yielding cumulative (downward plus sideline) EPNL and OASPL reductions of 7.7 dB and9.2 dB respectively. A fair correlation is established between the suppression of peakOASPL and the reduction of the radial velocity gradient on the underside of the jet.
Nomenclature
A Nozzle exit areaa Two-dimensional airfoil lift curve slopeCL Lift coefficientc Vane chord lengthDf Fan nozzle diameterf FrequencyG Radial velocity gradientJ ThrustL Vane liftM Mach numberp Pressurer Radial directionS Wedge wetted areaU Nozzle exit velocityu Mean axial velocity in jet plumew Average vane spanx Axial directiony Vertical directionz Horizontal transverse directionα Angle of attack, wedge half angleγ Specific heat ratioε Turning effortη Efficiencyθ Polar angleφ Azimuth angle
∗Professor, Department of Mechanical and Aerospace Engineering, 4200 Engineering Gateway, Irvine, CA 92697-3975, AIAAAssociate Fellow.
†Graduate Researcher, Department of Mechanical and Aerospace Engineering, 4200 Engineering Gateway, Irvine, CA 92697-3975, AIAA Student Member.
1 of 26
46th AIAA Aerospace Sciences Meeting and Exhibit7 - 10 January 2008, Reno, Nevada
This research builds upon the previous work performed at U.C. Irvine’s Jet Aeroacoustics Facility into theuse of fan flow deflector technology for supersonic jet noise reduction.1 The earlier studies used simple
nozzles and rather crude, external airfoil-type deflectors that are known to induce significant losses. Thecurrent study uses realistic nozzle shapes and leverages the advent of new types of deflectors, namely internalvanes and external wedges, developed in subsonic jet noise reduction efforts.2,3 Noise reduction is achievedin a coaxial separate-flow turbofan engine by tilting the bypass (secondary) plume downward by a smallamount relative to the core (primary) plume. The misalignment of the plumes creates a thick, low-speedregion on the underside of the jet which leads to a decrease in the convective Mach number of turbulenteddies in the jet shear layer. This principle, shown in Figure 1, leads to a reduction in the intense noise thatpropagates to the downward and sideline acoustic far-field.
Supersonic transportation is becoming increasingly viable both technologically and economically. How-ever, a major obstacle continues to be the problem of jet noise and the regulations governing the amountof noise that can be produced near airports. This research aims to develop innovative jet noise reductiontechniques for application on future civilian and military supersonic aircraft to facilitate environmental noisecompliance. The primary focus of the investigation is on the design, characterization and optimization ofmethods to reduce overall jet noise radiation towards airport communities. This is achieved through investi-gation of the interconnected elements of engine cycle, aerodynamic performance and noise reduction throughthe use of passive-control methods to asymmetrically reshape the exhaust plume. In this investigation fanflow deflectors are used to achieve mean flow distortion of the supersonic jet plume. The aircraft envisionedhas a cruise Mach number of 1.6 and weighs in the neighborhood of 9,000 kg. An engine cycle analysisfor the determination of optimal noise reduction is summarized and the results of acoustic experiments inconjunction with mean flow analysis of the jet plumes are presented to gain insight into an optimal fan flowdeflection configuration.
II. Engine Cycle Analysis
In this section we present a brief overview of the thermodynamic cycle analysis that leads to the de-termination of the exhaust conditions of the turbofan engine that are simulated in the experiment. Thebasic turbofan configuration shown in Figure 2 was analysed using a “conventional” gas turbine cycle withprovisions for turbine cooling and the ability to compare mixed-flow and unmixed-flow cycles. Table 1 sum-marizes the component efficiencies and specific heat ratios assumed. Figure 2 shows some of the principalvariables involved in the cycle analysis. The purpose of the analysis is to establish reasonable ranges forthe bypass ratio (BPR) and fan pressure ratio (FPR) of a state-of-the-art engine that would give acceptablecruise performance while ensuring compliance with takeoff noise regulations.
Figure 3 shows the results of such an exercise. It plots contour maps of thrust specific fuel consumption(TSFC) versus BPR and FPR at cruise (Mach 1.6) conditions for an unmixed-cycle engine with overallpressure ratio OPR = 15, rotor inlet temperature RIT = 1800 K, and turbine cooling mass flow ratioRcool = 0.2. The minimum TSFC occurs at BPR = 2.2 and FPR = 2.7. Such large FPR would require atwo-stage fan. Also, the BPR may be too small for noise compliance. Selecting the point BPR = 2.7, FPR= 2.2, shown by the red mark on Figure 3, mitigates these concerns without significant increase in TSFC. Ofcourse, the BPR cannot become too large without excessive drag penalties. Therefore, this research focuseson BPR between 2 and 3, and this specific report on BPR = 2.7.
Once the engine cycle parameters for cruise are chosen, they are then used to establish the takeoff exitconditions. The “Acoustic Tests column of Table 2 lists these conditions for BPR = 2.7, FPR = 2.2, andRIT =1800 K. The conditions are then reproduced in the subscale experiments.
2 of 26
III. Nozzle and Fan Flow Deflectors
The nozzle design in this report is based on the NASA GRC 3BB separate-flow nozzle, nominally forbypass ratio BPR = 5. The fan duct was reduced in diameter to produce BPR = 2.7 at the conditionsof the acoustics tests (Table 2), and the entire nozzle was scaled down by factor of eight to fit within theflow rate capability of the UCI facility. The nozzle design and exit coordinates of the nozzle are shown inFigures 4 and 5. The baseline nozzles were fabricated using a rapid prototyping epoxy method. The fan exitdiameter was Df=28.1 mm, and the fan exit height was 1.8 mm. The small fan exit height necessitated thefabrication of vanes of very small dimensions, of the same order as the fan exit height.
Fan flow deflection is achieved through the use of both internal airfoil-shaped vanes and external wedges/flapsas shown in Figure 6. The vanes were micro-machined from high-strength polycarbonate material usingCAD/CAM facilities at U.C. Irvine (Roland MDX-40 Subtractive Rapid Prototyping Milling machine). Thevane cross sections encompassed symmetric and asymmetric airfoils. The symmetric vanes have a NACA0012 airfoil cross section while the cambered vanes have a NACA 4412 cross section. The base and tip of eachvane are shaped to conform to the geometry of the fan and core ducts at the exact location where the vaneis attached. The vane chord length was 3 mm and the vane trailing edge was situated 2 mm upstream of thenozzle exit. Using one-dimensional theory it is estimated that the Mach numbers at the leading and trailingedges of the vane were 0.4 and 0.8, respectively. The external flaps had a length of 10 mm and half angleof 20 deg. They were positioned 5 mm downstream of the fan nozzle exit. The flaps were constructed froma fine metal mesh with a solidity of 50%. Parallel research efforts in high-bypass nozzles have shown thatperforated flaps (versus solid flaps or solid wedges) reduce velocity gradients in the vicinity of the deflector,thus reduce the potential of excessive noise created by the deflector itself.4
Nozzles were tested with both single-pair and two-pair vane configurations, with and without wedge-type flaps at various azimuthal angles and angles of attack. The particular configurations presented in thisdiscussion are displayed in Table 3. A four-vane configuration with vanes at azimuthal angles of 110 deg and165 deg and both with an angle of attack of 10 deg is shown in Figure 7.
A number of correlations can be made by investigating the total turning effort of the fan flow deflectorsand the effect on the jet noise produced. The turning effort is the total force of the fan flow deflectorsnormalized by the thrust of the bypass stream and can be viewed as a total deflection angle of the bypassstream from all of the deflectors. The turning effort is thus given by
ε =1Js
N∑
i=1
Li (1)
as a normalised sum of all of the individual vane and wedge lift forces. A computational study by Murayamaet al.5 for internal vane fan flow deflectors determined that the lift for the internal vane airfoil was the sameas an airfoil in external flow when the reference flow conditions were taken at the internal vane leading edge.The lift coefficient for the internal vane is then given by
CL = aαv (2)
where a is the 2-dimensional lift curve slope for the airfoil section used. Taking the flow conditions at thevane leading edge gives the total lift force
Lv = aαvqLEcw (3)
as a function of vane angle of attack, dynamic pressure at the vane leading edge qLE , vane chord lengthc and average vane span w. The aerodynamics of a wedge shaped fan flow deflector were investigated byPapamoschou et al.6 which determined the lift force of the wedge given by
Lw = CL(αw)qsS (4)
where the lift coefficient is dependent on the wedge half angle, the dynamic pressure is taken at the bypassnozzle exit and the wetted area S is determined by the wedge geometry and the material solidity.4
The individual lifts of each deflector can be applied in Equation 1 to give the total turning effort ε foreach fan flow deflector configuration.
3 of 26
IV. Aeroacoustic Testing
Aeroacoustic tests were conducted in U.C. Irvine’s Jet Aeroacoustics Facility, depicted in Figure 8. Thisis a subscale facility (approximately 1/40th of full scale for the tests in question) that uses helium-air mixturesfor simulating the exhaust velocity and density of hot jets.7 The exit flow conditions matched the conditionslisted in the “Acoustic Tests column of Table 2.
Jet noise was recorded by a microphone array consisting of eight 3.2 mm condenser microphones (Bruel& Kjaer, Model 4138) arranged on a circular arc centered at the vicinity of the nozzle exit. The polaraperture of the array is 30 deg and the array radius is 1 m. The angular spacing of the microphones islogarithmic. The entire array structure was rotated around its center to place the array at the desired polarangle. Positioning of the array is done remotely using a stepper motor. An electronic inclinometer displayedthe position of first microphone. Variations of the azimuth angle are possible by rotating the nozzle. Thisstudy encompassed the azimuth angles of 0 deg (downward) and 60 deg (sideline).
The arrangement of the microphones inside the anechoic chamber, and the principal electronic compo-nents, are shown in Figure 8. The microphones were connected, in groups of four, to two amplifier/signalconditioners (Bruel & Kjaer, Model 4138) with low-pass filter set at 300 Hz and high-pass filter set at 100kHz. The four-channel output of each amplifier was sampled at 250 kHz per channel by a multi-functiondata acquisition board (National Instruments PCI-6070E). Two such boards, one for each amplifier, wereinstalled in a Pentium 4 personal computer. National Instruments LabView software was used to acquirethe signals.
The sound pressure level spectrum was corrected for actuator response, free-field correction, and atmo-spheric absorption. The overall sound pressure level (OASPL) was obtained by integrating the correctedspectrum. The sound spectra are corrected for the frequency response of the microphone and for atmosphericabsorption. Through an elaborate procedure, the sound measurements are converted into effective perceivednoise level (EPNL) measured by ground observation points. A full discussion of the noise calculation processis presented by Papamoschou.2 The EPNL is evaluated for a constant-altitude flyover (1500-ft altitude) inthe downward (φ = 0 deg) and sideline (φ = 60 deg) directions. In this report the PNL and EPNL are basedon an engine thrust of 120.1 kN and an engine angle of attack of 10 deg. SPL spectra at full-scale size arepresented by dividing the laboratory frequencies by the scale factor of 44 and are referenced to a distance ofr/Df=1.25 from the jet nozzle exit.
V. Mean Velocity Surveys
Each acoustic test was followed by a mean velocity survey in a duplicate dual-stream apparatus. Insteadof helium-air mixtures, pure air was used in both primary and secondary streams. Therefore, the flowvelocities were lower than those in the acoustic tests. However, the velocity ratio Us/Up = 0.67, and primaryMach number Mp = 1.03, were held the same as in the acoustic tests. The Reynolds number of the jet,based on fan diameter, was 0.92 × 106 in the acoustic tests and 0.47 × 106 in the mean velocity surveys.
The mean axial velocity in the jet plume was surveyed using a Pitot rake system, shown in Figure 9. Therake consists of five 1-mm internal diameter Pitot probes attached to a three dimensional traverse system.The 70 mm long probes are spaced vertically 10 mm apart using a streamlined mounting plate. Each Pitotprobe is connected individually to a Setra Model 207 pressure transducer. The pressure was sampled at arate of 1000 Hz by an analog to digital data acquisition board (National Instruments PCI-MIO-16E). Machnumber and velocity were calculated from the Pitot pressure assuming constant pressure (equal to ambientvalue) and constant total temperature (equal to room temperature).
The three dimensional traverse system consists of three IMS MDrive 23 motor drivers connected indi-vidually to THK LM Guide Actuators. The traverse system is run remotely using National InstrumentsLabView over a pre-specified traverse array. The traverse array typically consisted of 28 axial planes span-ning 14 inches, each axial plane comprising 17 horizontal passes of length 101.6 mm spaced 2.5 mm verticallyapart. The horizontal passes were made at a speed of 10.16 mm/s.
The data on each y-z plane are interpolated on a fixed grid. The Pitot pressure is converted to velocityunder the assumption of constant static pressure (equal to the ambient value) and constant total temperature(equal to room temperature). Smoothing of the velocity profiles, and computation of the velocity gradients,is performed using a Savitzky-Golay filter.
For each axial station, the radial derivatives were calculated on the radial-azimuthal (r − φ) coordinate
4 of 26
system. The origin of the (r − φ) system is defined as the centroid of the region where the Pitot pressureexceeds 95% of its maximum value. The first and second derivatives were calculated along radial lines fromφ=0 to 354 deg in increments of 4 deg. The resulting radial velocity gradient is normalized in the form
G(x, r, φ) =Df
Up
∂u(x, r, φ)∂r
Of particular interest is the maximum value of the magnitude of the gradient for given x and φ,
Gmax(x, φ) =Df
Up
∣∣∣∣∂u(x, r, φ)
∂r
∣∣∣∣max
VI. Aeroacoustic Results
Aeroacoustic results are presented with respect to the baseline nozzle results. Each figure shows narrow-band lossless spectra at various polar angles θ (measured from the jet axis), overall sound pressure (OASPL)versus θ, perceived noise level (PNL) versus time, and PNL versus polar angle. The reductions in EPNL andpeak value of OASPL are also shown. These aeroacoustic attributes facilitate a determination of the merit ofthe various fan flow deflector configurations. As stated previously, 2-, 4- and 6-vane and wedge combinationconfigurations were investigated with both symmetric and cambered vanes and mesh wedges. A number ofvarious configurations are now presented to illuminate the process of the fan flow deflector optimization.
The aeroacoustic results for a configuration consisting of a single pair of symmetric NACA 0012 airfoilvanes at an azimuth angle φ = 150◦ and an angle of attack α = 7.5◦ are shown in Figure 10 and 11for microphone angles 0 deg and 60 deg respectively as measured from the vertical downward direction.This configuration results in very small downward EPNL and peak-OASPL reductions of 0.4 dB and 0.7dB respectively and marginally better sideline EPNL and peak-OASPL reductions of 1.2 dB and 1.7 dBrespectively. These tests configuration indicate that a small pair of vanes, of the small size used here(approximately 2 × 3 mm) may not be sufficient to provide significant acoustic reductions.
When an additional pair of symmetric vanes is added a number of changes in the jet noise are realized.Figures 12 and 13 show the aeroacoustic results for a 4-vane configuration with a pair of symmetric NACA0012 vanes at φ = 150◦ and α = 7.5◦ as before with another pair of symmetric vanes placed at an azimuthangle φ = 90◦ with an angle of attack also set at α = 7.5◦. The downward EPNL and peak-OASPLreductions are improved to 3.4 dB and 3.5 dB respectively. The sideline EPNL and peak-OASPL reductionsare improved to 3.1 dB and 3.2 dB, respectively. One deficiency of this configuration is the higher OASPLlevels seen at larger polar angles.
Further improvements are seen when cambered vanes are used, which are able to give superior lift-to-drag ratios over the symmetric vanes. Figure 14 and 15 show the aeroacoustic results for a promising 4-vaneconfiguration for microphone angles of 0 deg and 60 deg respectively. This configuration consists of 2 pairsof micromachined vanes with a NACA 4412 cambered airfoil cross section. The top pair of vanes was setat an azimuth angle φ = 150 deg with an angle of attack α = 4 deg. The bottom pair were set at φ = 90deg and α = 7.5 deg. This configuration results in good downward EPNL and peak-OASPL reductionsof 3 dB and 3.3 dB respectively and very good sideline EPNL and peak-OASPL reductions of 4.1 dB and3.6 dB respectively. This particular configuration has encouraging attributes as it gives significant spectralreductions at lower polar angles resulting in significant OASPL reductions together with negligible spectralincreases at higher polar angles.
A configuration consisting of a single external mesh wedge as shown in Figure 16 gives the acoustic resultsin Figure 17 and 18 for the downward and sideline microphone angles respectively. Very good downwardEPNL and OASPL reductions of 3.9 dB and 4.1 dB respectively are realized whilst giving moderate 2.5 dBand 2.3 dB reductions in the sideline EPNL and OASPL respectively.
A configuration combining two pairs of cambered vanes and a mesh wedge results in further reductionsin EPNL and OASPL at both downward and sideline microphone angles. The assembled mesh wedgeand micromachined vane configuration is shown in Figure 19. The corresponding acoustic results for thedownward and sideline microphone angles are shown in Figure 20 and 21. This configuration results indownward EPNL and peak-OASPL reductions of 4.4 dB and 5.2 dB respectively and sideline EPNL andpeak-OASPL reductions of 3.3 dB and 4.0 dB respectively. While there is a minor decrease in the sidelineEPNL reduction from the 4-vane configuration due to a slight increase in the spectral levels at higher polar
5 of 26
angles, this is offset by further sideline OASPL reductions. The downward reductions in EPNL and OASPLare very promising and with refinements it is hoped to produce EPNL and OASPL reductions of over 5 dBfor both downward and sideline measurements.
In addition to conducting spectral analyses, it is important to also examine statistics in the time domain.The skewness of the fluctuating pressure signal can be utilized to further analyze the reductions evident inall fan flow deflection acoustic measurements. The skewness is defined as
Skewness =p′3
p′32
(5)
and has been shown by Ffowcs-Williams et al.8 to have direct correlations to the emission of Mach wavesand accompanying “crackle” noise. A decrease in the skewness of the pressure signal indicates a relativesuppression of Mach wave radiation and hence decreased levels of noise radiated to the far field. Timetracesof the pressure signal for the baseline and 4V+W case are shown in Figure 22 for the polar angle θ = 45 deg.The resulting skewness of the timetraces are calculated according to Equation 5 and are shown in Figure 23.The large reductions in the peak level of skewness of case 4V+W, almost 50% below the baseline, indicatesa significant reduction in the emission of Mach waves.
Results from all the acoustic experiments have been analysed to determine the underlying trends inorder to find the best configuration. The turning effort of each configuration was determined accordingto Murayama et al.5 and compared to the cumulative maximum EPNL and OASPL reductions from bothdownward and sideline microphone positions measured in the aeroacoustics facility. The results are presentedin Figure 24. A general trend is evident that suggests a positive correlation between turning effort andEPNL and OASPL reduction. It was observed however that placing both the symmetric and camberedmicromachined vanes at angles of attack larger than α = 7.5 deg resulting in increased noise levels in thebroadside direction.
While a general trend of better sound reduction with increased turning effort is present a compromisemust be reached due to possible increases in drag especially when the wedge is implemented. It is envisagedthat the wedge will be deployable and only used during takeoff when jet noise reduction is required.
VII. Mean Velocity Results
The results for the mean flow of the jets are presented as a composite of the velocity contour for thevertical longitudinal (symmetry) plane, a number of transverse velocity contour plots at various distancesdownstream and several velocity line plots on the symmetry plane. Figure 25 and 26 display the mean velocityresults and the maximum-gradient Gmax(x, φ) contours for the baseline nozzle. The velocity isocontours inthe two planes presented are fairly axisymmetric and the maximum gradient contours are relatively constantindicating that the nozzle components (fan, core and plug) are in good alignment. The line plots alsoconfirm this symmetry in the vertical plane. The primary potential core length, xp, is defined as the lengthdownstream of the plug tip at which the local maximum velocity falls to 90% of the core exit velocity. The endof the potential core is of interest as it has been identified from phased array measurements that the strongestnoise sources originate from this region.9 For the baseline case the potential core length xp/Df = 4.5.
The mean flow results for the configuration consisting of a single pair of NACA 0012 airfoil vanes aredisplayed in Figure 27 corresponding to the acoustic results in Figures 10 and 11. The addition of the vanescauses distorted cross-sections with thickening mainly in the lateral (φ = 90 deg) direction. A decrease in thepotential core length to xp/Df = 3.3 also occurs when the vanes are added. Figure 28 displays contours ofmaximum radial velocity gradient on the x− φ plane and the relative change in maximum velocity gradientagainst the baseline case. It is seen that most of the gradient reduction occurs at φ = ±90 deg. Thismay explain the poor noise reduction in the downward direction and the moderate reduction in the sidelinedirection. It should also be noted that this 2V configuration employed vanes manufactured in the earlystages of our micromachining effort that were not as refined as later versions. This may have produced a“dirtier” flow in the vicinity of the nozzle exit.
The mean flow results for case 4V with symmetric vanes are shown in Figure 29 corresponding to theacoustic results of Figures 12 and 13. The longitudinal cross section plot shows a potential core lengthxp/Df = 4.1, which is also reduced from the baseline case. The transverse cross-section plots show thethickening of the plume on the underside of the jet, which displays a thicker region than the 2-vane config-uration in the downward direction. This can also be seen in Figure 30 which shows the maximum velocity
6 of 26
gradients and relative change in maximum velocity gradients for this configuration. In the vicinity of thenozzle, the gradients are reduced mostly at φ = ±60 deg. However, near the end of the potential core(x/Df ∼ 5) there is a uniform reduction in the gradient for −60 deg ≤ φ ≤ 60 deg. This is consistent withthe roughly-equal noise reductions in the downward and sideline reductions seen in Figures 12 and 13.
The use of cambered airfoils in the 4V configuration gives the mean flow results shown in Figure 31.The corresponding acoustic results are in Figures 14 and 15. In comparison to the 4V case with symmetricvanes (Figure 29), the velocity contours are similar but slightly thickened in the lateral direction. Indeedthe maximum velocity gradient contour plots, Figure 32, show a slightly larger decrease in gradient atazimuth angles in the vicinity of ±60 deg which is consistent with the better sideline acoustic results of thisconfiguration.
The effects of using the mesh flaps as a fan flow deflector are evident in the mean flow results in Figure33, which correspond to the acoustic results in Figures 17 and 18. A strong deflection is evident in thevicinity of the wedge and a thickening of the plume develops either side of the wedge on the top of thejet. This region migrates around the jet several fan diameters downstream to form a thicker region on theunderside of the jet which results in the noise reductions presented in the previous section that are betterin the downward direction than in the sideline direction. The potential core length for this configuration isxp/Df = 4.2, a small reduction from the baseline. The reduction in maximum velocity gradient, which canbe seen in Figure 34, is reasonably uniform over all azimuth angles except near the top of the jet (φ = 180deg) where strong gradients are evident very close to the nozzle exit.
The configuration consisting of both the mesh wedge and 2 pairs of cambered vanes (4V+W), correspond-ing to the acoustic results in Figures 20 and 21 produces the mean flow results shown in Figure 35. The samevelocity contour features from the previous two configurations are evident, with a large low speed region onthe underside of the jet as a result of both the cambered vanes and the displaced flow from the mesh flaps.A decreased potential core length from the previous two configurations xp/Df = 3.8 is realized. Figure 36indicates that significant reductions in the maximum velocity gradient are concentrated in the downwarddirection in the vicinity of the end of the potential core, consistent with the noise reductions measured inthe acoustic experiments.
One of the goals of this research is to correlate changes in the mean velocity, induced by the fan flowdeflectors, the corresponding changes in acoustics. The maximum velocity gradient is evidently an importantmean flow parameter that has implications for jet noise reduction. Figure 37 shows the correlation betweenthe maximum gradient measured in the downward and sideline directions and the reduction in peak OASPLin the same directions. The maximum velocity gradient was averaged over 2.7 to 5.4 fan nozzle diametersdownstream of the plug tip. A general negative trend is realized indicating the importance of producing athick low speed region around a large range of azimuth angles on the underside of the jet plume in orderto give significant acoustic reductions in both downward and sideline directions. The trends in Figure37 represent our first attempt to correlate noise with flow gradients and in the future we will be seekingimprovements in those correlations.
VIII. Conclusion
We report the initial phase of our research on fan flow deflectors for jet noise reduction in next-generationsupersonic turbofan engines. An engine cycle analysis was performed to determine reasonable ranges for thebypass ratio, fan pressure ratio, and takeoff exhaust conditions. A cycle with BPR=2.7 and FPR=2.2was selected. The parametric investigation of fan flow deflectors used acoustic experiments in conjunctionwith mean velocity surveys to determine the efficacy of several different deflector technologies. Symmetricand cambered micromachined airfoil internal vanes were employed. In addition, an external wedge-typeperforated flap deflector was tried. Significant reductions were achieved with all methods however severalpertinent trends were realized:
• Internal vanes gave superior noise reduction when they were placed at azimuth angles greater than orequal to 90 deg, which produces thick low speed region on the underside of the jet that extends to thedownward direction and sideline direction.
• Two pairs of internal vanes produced a more uniform low speed region on the underside of the jetplume compared to a single pair of vanes. This resulted in superior noise reduction in both downwardand sideline directions.
7 of 26
• Vanes with a cambered airfoil section gave superior acoustic results over those with a symmetric airfoilsection possibly due to the higher lift to drag ratios of the cambered airfoil. A smaller vane angle ofattack can be used for the cambered airfoil giving a lower relative blockage of the fan duct.
• A deployable wedge with a solidity of 0.5 gave significant noise reductions, particularly in the downwarddirection and produced a uniform low speed region on the underside of the jet.
• Superior noise reductions in both the downward and sideline directions was achieved with a combinationof two pairs of cambered airfoil vanes and a mesh wedge. This configuration resulted in cumulativeEPNL and OASPL reductions of 7.7 dB and 9.2 dB respectively.
• A general positive trend between deflector turning effort and noise reduction was found. In addition,there is a fair correlation between the reduction in peak OASPL and the reduction of the maximumradial velocity gradient on the underside of the jet near the end of the primary potential core.
Acknowledgments
The support by NASA Cooperative Agreement NNX07AC62A (monitored by Mr. Tom Norum) is grate-fully acknowledged. We thank Mr. An Vu for his assistance with the design and fabrication of the vanes.
References
1Papamoschou, D., “Engine Cycle Exhaust Configuration for Quiet Supersonic Propulsion,” AIAA Journal of Propulsionand Power , Vol. 20, No. 2, 2004, pp. 255–262.
2Papamoschou, D., “Fan Flow Deflection in Simulated Turbofan Exhaust,” AIAA Journal , Vol. 44, No. 12, 2006, pp. 3088–3097.
3Zaman, K., Bridges, J., and Papamoschou, D., “Offset Stream Technology - Comparison of Results from UCI and GRCExperiments,” AIAA Paper 2007-0438, January 2007.
4Papamoschou, D., “Pylon Based Fan Flow Deflectors,” AIAA Paper 2008-0040, January 2008.5Murayama, T., Papamoschou, D., and Liu, F., “Aerodynamics of Fan flow Deflectors for Jet Noise Suppression,” AIAA
Paper 2005-0994, January 2005.6Papamoschou, D., Vu, A., and Johnson, A., “Aerodynamics of Wedge-Shaped Deflectors for Jet Noise Reduction,” AIAA
Paper 2006-3655, June 2006.7Papamoschou, D., “Acoustic Simulation of Hot Coaxial Jets Using Cold Helium-Air Mixture Jets,” AIAA Journal of
Propulsion and Power , Vol. 23, No. 2, 2007, pp. 375–381.8Ffowcs-Williams, J., Simson, J., and Virchis, V., “Crackle: an Annoying Component of Jet Noise,” Journal of Fluid
Mechanics, Vol. 71, No. 2, 1975, pp. 251–271.9Narayanan, S., Barber, T., and Polak, D., “High Subsonic Jet Experiments: Turbulence and Noise Generation Studies,”
AIAA Journal , Vol. 40, No. 3, 2002, pp. 430–437.
Core plumeBypass plume
Deflectors
Figure 1. General concept of fan flow deflection.
8 of 26
1 2 8 3 4 5 9 74L4H
Overall Pressure Ratio OPR
Fan Pressure Ratio FPR
Rotor Inlet Temperature
RIT
Bypass RatioBPR
Figure 2. Turbofan engine model used in the thermodynamic cycle analysis.
Table 1. Component efficiencies and specific heat ratios assumed for the BPR = 2.7 engine.
Figure 3. Example result of cycle analysis for Mach 1.6 cruise. Contour maps of thrust specific fuel consumptionversus bypass ratio and fan pressure ratio. Red dot indicates design selected.
Figure 6. Illustration of internal vane (left) and external wedge/flap (right) fan flow deflectors.
Figure 7. Front and top view (without the fan nozzle) of a 4-vane nozzle configuration.
11 of 26
Two Nexus 2690-A-OS4Conditioning Amplifiers
8 BK-4138Microphones
PC with two PCI-6070E 1.2 MS/sDAQ boards
Compressedair and helium mixtures
Anechoic chamber 1.9 x 2.2 x 2.2 m
Two SCB-68Blocks
Jet nozzle
Circular arc path
Two Nexus 2690-A-OS4Conditioning Amplifiers
8 BK-4138Microphones
PC with two PCI-6070E 1.2 MS/sDAQ boards
Compressedair and helium mixtures
Anechoic chamber 1.9 x 2.2 x 2.2 m
Two SCB-68Blocks
Jet nozzle
Circular arc path
8 BK-4138Microphones
PC with two PCI-6070E 1.2 MS/sDAQ boards
Compressedair and helium mixtures
Anechoic chamber 1.9 x 2.2 x 2.2 m
Two SCB-68Blocks
Jet nozzle
Circular arc path
Figure 8. UCI Jet Aeroacoustics Facility.
x
y
z x
y
z1.0 mm ID
10 mm
Figure 9. The Pitot traverse system in the Supersonic Mean Flow Facility.
12 of 26
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M101B3M108
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M101B3M108
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M101B3M108
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M101B3M108
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M101B3M108
0 20 40 60 80 100 120130
135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M101B3M108
40 45 50 55 6070
80
90
100
110PN
L(dB
)
Time (sec)
B3M101 EPNL=102.0B3M108 EPNL=101.7
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB0.4OASPL
peak= dB0.7
(Positive = reduction)
B3M101 EPNL=102.0B3M108 EPNL=101.7
Figure 10. Acoustic results for symmetric 2-vane configuration (2V) with comparison to baseline. Microphoneazimuth angle φ = 0 deg (downward).
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M101B3M109
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M101B3M109
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M101B3M109
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M101B3M109
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M101B3M109
0 20 40 60 80 100 120130
135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M101B3M109
40 45 50 55 6070
80
90
100
110
PNL(
dB)
Time (sec)
B3M101 EPNL=102.0B3M109 EPNL=100.8
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB1.2OASPL
peak= dB1.7
(Positive = reduction)
B3M101 EPNL=102.0B3M109 EPNL=100.8
Figure 11. Acoustic results for symmetric 2-vane configuration (2V) with comparison to baseline. Microphoneazimuth angle φ = 60 deg (sideline).
13 of 26
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M091B3M124
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M091B3M124
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M091B3M124
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M091B3M124
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M091B3M124
0 20 40 60 80 100 120135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M091B3M124
40 45 50 55 6070
80
90
100
110PN
L(dB
)
Time (sec)
B3M091 EPNL=104.4B3M124 EPNL=101.1
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB3.4OASPL
peak= dB3.6
(Positive = reduction)
B3M091 EPNL=104.4B3M124 EPNL=101.1
Figure 12. Acoustic results for symmetric 4-vane configuration (4V) with comparison to baseline. Microphoneazimuth angle φ = 0 deg (downward).
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M091B3M125
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M091B3M125
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M091B3M125
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M091B3M125
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M091B3M125
0 20 40 60 80 100 120135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M091B3M125
40 45 50 55 6070
80
90
100
110
PNL(
dB)
Time (sec)
B3M091 EPNL=104.4B3M125 EPNL=101.3
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB3.1OASPL
peak= dB3.2
(Positive = reduction)
B3M091 EPNL=104.4B3M125 EPNL=101.3
Figure 13. Acoustic results for symmetric 4-vane configuration (4V) with comparison to baseline. Microphoneazimuth angle φ = 60 deg (sideline).
14 of 26
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M091B3M142
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M091B3M142
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M091B3M142
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M091B3M142
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M091B3M142
0 20 40 60 80 100 120135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M091B3M142
40 45 50 55 6070
80
90
100
110PN
L(dB
)
Time (sec)
B3M091 EPNL=104.4B3M142 EPNL=101.4
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB3.0OASPL
peak= dB3.3
(Positive = reduction)
B3M091 EPNL=104.4B3M142 EPNL=101.4
Figure 14. Acoustic results for cambered 4-vane configuration (4V) with comparison to baseline. Microphoneazimuth angle φ = 0 deg (downward).
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M091B3M143
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M091B3M143
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M091B3M143
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M091B3M143
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M091B3M143
0 20 40 60 80 100 120135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M091B3M143
40 45 50 55 6070
80
90
100
110
PNL(
dB)
Time (sec)
B3M091 EPNL=104.4B3M143 EPNL=100.4
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB4.1OASPL
peak= dB3.6
(Positive = reduction)
B3M091 EPNL=104.4B3M143 EPNL=100.4
Figure 15. Acoustic results for cambered 4-vane configuration (4V) with comparison to baseline. Microphoneazimuth angle φ = 60 deg (sideline).
15 of 26
Figure 16. Mesh wedge configuration.
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M091B3M144
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M091B3M144
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M091B3M144
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M091B3M144
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M091B3M144
0 20 40 60 80 100 120135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M091B3M144
40 45 50 55 6070
80
90
100
110
PNL(
dB)
Time (sec)
B3M091 EPNL=104.4B3M144 EPNL=100.5
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB3.9OASPL
peak= dB4.1
(Positive = reduction)
B3M091 EPNL=104.4B3M144 EPNL=100.5
Figure 17. Acoustic results for mesh wedge configuration (W) with comparison to baseline. Microphoneazimuth angle φ = 0 deg (downward).
16 of 26
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M091B3M145
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M091B3M145
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M091B3M145
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M091B3M145
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M091B3M145
0 20 40 60 80 100 120135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M091B3M145
40 45 50 55 6070
80
90
100
110PN
L(dB
)
Time (sec)
B3M091 EPNL=104.4B3M145 EPNL=101.9
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB2.5OASPL
peak= dB2.3
(Positive = reduction)
B3M091 EPNL=104.4B3M145 EPNL=101.9
Figure 18. Acoustic results for mesh wedge configuration (W) with comparison to baseline. Microphoneazimuth angle φ = 60 deg (sideline).
Figure 19. Cambered 4-vane and mesh wedge configuration.
17 of 26
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M091B3M152
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M091B3M152
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M091B3M152
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M091B3M152
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M091B3M152
0 20 40 60 80 100 120135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M091B3M152
40 45 50 55 6070
80
90
100
110PN
L(dB
)
Time (sec)
B3M091 EPNL=104.4B3M152 EPNL=100.0
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB4.4OASPL
peak= dB5.2
(Positive = reduction)
B3M091 EPNL=104.4B3M152 EPNL=100.0
Figure 20. Acoustic results for cambered 4-vane and mesh wedge configuration (4V+W) with comparison tobaseline. Microphone azimuth angle φ = 0 deg (downward).
0.02 0.1 0.2 0.5 1 2859095
100105110115120125130
SPL(
dB/H
z)
f (kHz) - full scale
Spectra at =30o
B3M091B3M153
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
120
125
f (kHz) - full scale
Spectra at =50o
B3M091B3M153
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
115
f (kHz) - full scale
Spectra at =70o
B3M091B3M153
0.02 0.1 0.2 0.5 1 285
90
95
100
105
110
f (kHz) - full scale
Spectra at =90o
B3M091B3M153
0.02 0.1 0.2 0.5 1 285
90
95
100
105
f (kHz) - full scale
Spectra at =120o
B3M091B3M153
0 20 40 60 80 100 120135
140
145
150
155
160
OA
SPL(
dB)
(deg)
B3M091B3M153
40 45 50 55 6070
80
90
100
110
PNL(
dB)
Time (sec)
B3M091 EPNL=104.4B3M153 EPNL=101.1
02040608010012014070
80
90
100
110
PNL(
dB)
Polar angle , deg
EPNL= dB3.3OASPL
peak= dB4.0
(Positive = reduction)
B3M091 EPNL=104.4B3M153 EPNL=101.1
Figure 21. Acoustic results for cambered 4-vane and mesh wedge configuration (4V+W) with comparison tobaseline. Microphone azimuth angle φ = 60 deg (sideline).
18 of 26
-300
-200
-100
0
100
200
300
0 2 4 6 8 10Time (ms)
Pre
ssur
e (P
a)
-300
-200
-100
0
100
200
300
0 2 4 6 8 10Time (ms)
Pre
ssur
e (P
a)
Figure 22. Timetraces of the microphone pressure for the baseline (left) and 4-vane and wedge (right) config-urations at θ = 40 deg.
0 20 40 60 80 100 1200
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Skew
ness
(deg)
B3M091 (max=0.65)B3M152 (max=0.36)
Figure 23. The directivity of the skewness for the baseline and 4-vane plus wedge configurations.
0.00
2.00
4.00
6.00
8.00
10.00
0.00 0.50 1.00 1.50 2.00 2.50Turning Effort (deg)
OA
SP
Lcum
0.00
2.00
4.00
6.00
8.00
10.00
0.00 0.50 1.00 1.50 2.00 2.50Turning Effort (deg)
EP
NLc
um
Figure 24. The effect of turning effort on cumulative OASPL (left) and cumulative EPNL (right).
19 of 26
x/Df
y/D
f
B3P100 Contours of u(x,y,0)/Up
0 1 2 3 4 5 6 7 8 9 10 11 12
-1
-0.5
0
0.5
1
0.10.20.30.40.50.60.70.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
y/D
f
u(y) at x/Df=0
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=1.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=3.7
z/Df
y/D
f
u(y,z)/Up at x/Df=0
-1 -0.5 0 0.5 1-1
-0.5
0
0.5
1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=1.8
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=3.6
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=5.4
-1 -0.5 0 0.5 1
0.2
0.4
0.6
z/Df
u(y,z)/Up at x/Df=7.2
-1 -0.5 0 0.5 1
0.2
0.4
0.6
Figure 25. Mean flow results for baseline nozzle.
x/Df
(deg
)
Contours of maximum radial velocity gradient
0 2 4 6 8 10 12-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
0.5
1
1.5
2
2.5
3
Figure 26. Contours of maximum velocity gradient on x − φ plane for baseline jet.
20 of 26
x/Df
y/D
f
B3P108 Contours of u(x,y,0)/Up
0 1 2 3 4 5 6 7 8 9 10 11 12
-1
-0.5
0
0.5
1
0.10.20.30.40.50.60.70.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
y/D
f
u(y) at x/Df=0
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=1.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=3.7
z/Df
y/D
f
u(y,z)/Up at x/Df=0
-1 -0.5 0 0.5 1-1
-0.5
0
0.5
1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=1.8
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=3.6
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=5.4
-1 -0.5 0 0.5 1
0.2
0.4
0.6
z/Df
u(y,z)/Up at x/Df=7.2
-1 -0.5 0 0.5 1
0.1
0.2
0.3
0.4
0.5
Figure 27. Mean velocity results for symmetric 2-vane configuration (2V).
x/Df
(deg
)
Contours of maximum radial velocity gradient
0 2 4 6 8 10 12-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
0.5
1
1.5
2
2.5
3
3.5
x/Df
(deg
)
Relative change in radial velocity gradient
0 1 2 3 4 5 6 7 8 9 10-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
Figure 28. Contours of maximum radial velocity gradient on x−φ plane shown in absolute (left) and differential(right) forms for symmetric 2-vane configuration (2V).
21 of 26
x/Df
y/D
f
B3P124 Contours of u(x,y,0)/Up
0 1 2 3 4 5 6 7 8 9 10 11 12
-1
-0.5
0
0.5
1
0.10.20.30.40.50.60.70.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
y/D
f
u(y) at x/Df=0
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=1.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=3.7
z/Df
y/D
f
u(y,z)/Up at x/Df=0
-1 -0.5 0 0.5 1-1
-0.5
0
0.5
1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=1.8
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=3.6
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=5.4
-1 -0.5 0 0.5 1
0.2
0.4
0.6
z/Df
u(y,z)/Up at x/Df=7.2
-1 -0.5 0 0.5 1
0.1
0.2
0.3
0.4
0.5
Figure 29. Mean velocity results for symmetric 4-vane configuration (4V).
x/Df
(deg
)
Contours of maximum radial velocity gradient
0 2 4 6 8 10 12-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
0.5
1
1.5
2
2.5
3
3.5
x/Df
(deg
)
Relative change in radial velocity gradient
0 1 2 3 4 5 6 7 8 9 10-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
Figure 30. Contours of maximum radial velocity gradient on x−φ plane shown in absolute (left) and differential(right) forms for symmetric 4-vane configuration (4V).
22 of 26
x/Df
y/D
f
B3P142 Contours of u(x,y,0)/Up
0 1 2 3 4 5 6 7 8 9 10 11 12
-1
-0.5
0
0.5
1
0.10.20.30.40.50.60.70.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
y/D
f
u(y) at x/Df=0
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=1.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=3.7
z/Df
y/D
f
u(y,z)/Up at x/Df=0
-1 -0.5 0 0.5 1-1
-0.5
0
0.5
1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=1.8
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=3.6
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=5.4
-1 -0.5 0 0.5 1
0.2
0.4
0.6
z/Df
u(y,z)/Up at x/Df=7.2
-1 -0.5 0 0.5 1
0.1
0.2
0.3
0.4
0.5
Figure 31. Mean velocity results for cambered 4-vane configuration (4V).
x/Df
(deg
)
Contours of maximum radial velocity gradient
0 2 4 6 8 10 12-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
0.5
1
1.5
2
2.5
3
x/Df
(deg
)
Relative change in radial velocity gradient
0 1 2 3 4 5 6 7 8 9 10-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
Figure 32. Contours of maximum radial velocity gradient on x−φ plane shown in absolute (left) and differential(right) forms for cambered 4-vane configuration (4V).
23 of 26
x/Df
y/D
f
B3P144 Contours of u(x,y,0)/Up
0 1 2 3 4 5 6 7 8 9 10 11 12
-1
-0.5
0
0.5
1
0.10.20.30.40.50.60.70.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
y/D
f
u(y) at x/Df=0
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=1.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=3.7
z/Df
y/D
f
u(y,z)/Up at x/Df=0
-1 -0.5 0 0.5 1-1
-0.5
0
0.5
1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=1.8
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=3.6
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=5.4
-1 -0.5 0 0.5 1
0.2
0.4
0.6
z/Df
u(y,z)/Up at x/Df=7.2
-1 -0.5 0 0.5 1
0.10.20.30.40.5
Figure 33. Mean velocity results for mesh wedge configuration (W).
x/Df
(deg
)
Contours of maximum radial velocity gradient
0 2 4 6 8 10 12-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
0.5
1
1.5
2
2.5
3
3.5
x/Df
(deg
)
Relative change in radial velocity gradient
0 1 2 3 4 5 6 7 8 9 10-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
Figure 34. Contours of maximum radial velocity gradient on x−φ plane shown in absolute (left) and differential(right) forms for mesh wedge configuration (W).
24 of 26
x/Df
y/D
f
B3P152 Contours of u(x,y,0)/Up
0 1 2 3 4 5 6 7 8 9 10 11 12
-1
-0.5
0
0.5
1
0.10.20.30.40.50.60.70.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
y/D
f
u(y) at x/Df=0
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=1.8
0 0.5 1-1
-0.5
0
0.5
1
u(y)/Up
u(y) at x/Df=3.7
z/Df
y/D
f
u(y,z)/Up at x/Df=0
-1 -0.5 0 0.5 1-1
-0.5
0
0.5
1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=1.8
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=3.6
-1 -0.5 0 0.5 1
0.2
0.4
0.6
0.8
z/Df
u(y,z)/Up at x/Df=5.4
-1 -0.5 0 0.5 1
0.2
0.4
0.6
z/Df
u(y,z)/Up at x/Df=7.2
-1 -0.5 0 0.5 1
0.1
0.2
0.3
0.4
0.5
Figure 35. Mean velocity results for cambered 4-vane and mesh wedge configuration (4V+W).
x/Df
(deg
)
Contours of maximum radial velocity gradient
0 2 4 6 8 10 12-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
0.5
1
1.5
2
2.5
3
3.5
4
x/Df
(deg
)
Relative change in radial velocity gradient
0 1 2 3 4 5 6 7 8 9 10-180-160-140-120-100
-80-60-40-20
020406080
100120140160180
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
Figure 36. Contours of maximum radial velocity gradient on x−φ plane shown in absolute (left) and differential(right) forms for cambered 4-vane and mesh wedge configuration (4V+W).
25 of 26
0
1
2
3
4
5
6
1.2 1.4 1.6 1.8 2.0 2.2Maximum Gradient (ave.)
OA
SP
Lmax
DownwardSideline
Figure 37. Correlation of suppression in peak OASPL with reduction of maximum radial velocity gradientGmax averaged over the axial extent 2.7 < x/Df < 5.4.
